CN103294746A - Method and device of de-identification in visual media data - Google Patents
Method and device of de-identification in visual media data Download PDFInfo
- Publication number
- CN103294746A CN103294746A CN2013100163589A CN201310016358A CN103294746A CN 103294746 A CN103294746 A CN 103294746A CN 2013100163589 A CN2013100163589 A CN 2013100163589A CN 201310016358 A CN201310016358 A CN 201310016358A CN 103294746 A CN103294746 A CN 103294746A
- Authority
- CN
- China
- Prior art keywords
- image
- average image
- word
- average
- visual media
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/62—Protecting access to data via a platform, e.g. using keys or access control rules
- G06F21/6218—Protecting access to data via a platform, e.g. using keys or access control rules to a system of files or objects, e.g. local or distributed file system or database
- G06F21/6245—Protecting personal data, e.g. for financial or medical purposes
- G06F21/6254—Protecting personal data, e.g. for financial or medical purposes by anonymising data, e.g. decorrelating personal data from the owner's identification
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
Abstract
A method of de-identification in visual media data is described. The method includes merging a sequence of images from a set of visual media data into an averaged image; bounding portions of the averaged image that are determined to be relatively fixed, wherein each bounded portion is identified by a corresponding position in the averaged image; generating a template comprising the bounded portions and the corresponding position for each bounded portion in the averaged image; and de-identifying the sequence of images by obfuscating content in the bounded portions.
Description
Technical field
Control demand and professional consideration usually require the mode swap data with safety, particularly in health care." health insurance portability and accountability act " in 1996 (HIPAA) stipulated to require safely swap data and non-disclosure patient information specific.Therefore, the data of some type must be modified with obfuscation (obfuscate) sensitivity or security information before by exchange.
Solution goes the method for identification (de-identification) to address the problem usually: what (1) should go to identify, (2) how much should go identification, should and how (3) go identification? the several different methods of going to identify document and metadata comprises that removal is marked as the built-in code of the part of identifying, and perhaps writes the method based on template from the information of the document in the document classification.Go recognition methods be used to text document, such as the structural metadata field in digital imaging and communications in medicine (DICOM) metadata, but the visual media data when sign embeds in the content to go to identify be difficult with consuming time.
Summary of the invention
The embodiment of system has been described.In one embodiment, this system is that visual media goes recognition system.This system comprises: the image combiner, and it is configured to and will merges into the average image from the image sequence of visual media data acquisition; And removing to identify engine, it is configured to: define the relatively-stationary part that is confirmed as of (bound) the average image, wherein each is defined part and is identified by the relevant position in the average image; Generation comprises that each that be defined part and be used for the average image is defined the template of the relevant position of part; And by obfuscation be defined the part content make a return journey identification described image sequence.Other embodiment of this system have also been described.
The embodiment of computer program has also been described.In one embodiment, this computer program comprises the computer-readable recording medium of storing computer-readable program, and wherein this computer-readable program makes this computing machine carry out the operation that is used for going to identify the visual media data when being carried out by the processor in the computing machine.Described operation comprises: will merge into the average image from the image sequence of visual media data acquisition; Define the relatively-stationary part that is confirmed as of the average image, wherein each is defined part and is identified by the relevant position in the average image; Generation comprises that each that be defined part and be used for the average image is defined the template of the relevant position of part; And by obfuscation be defined the part content make a return journey identification described image sequence.Other embodiment of this device have also been described.
The embodiment of method has also been described.In one embodiment, this method is be used to the method for going to identify the visual media data.This method comprises: will merge into the average image from the image sequence of visual media data acquisition; Define the relatively-stationary part that is confirmed as of the average image, wherein each is defined part and is identified by the relevant position in the average image; Generation comprises that each that be defined part and be used for the average image is defined the template of the relevant position of part; And by obfuscation be defined the part content make a return journey identification described image sequence.Other embodiment of this method have also been described.
Description of drawings
Fig. 1 shows the synoptic diagram that visual media removes an embodiment of recognition system;
The visual media that Fig. 2 shows Fig. 1 removes the synoptic diagram of an embodiment of recognition system;
Fig. 3 shows for generating the process flow diagram of an embodiment that visual media goes the method for recognition template;
Fig. 4 shows the synoptic diagram of an embodiment of the average image of Fig. 1;
Fig. 5 shows the synoptic diagram of an embodiment of the average image of Fig. 1;
The visual media that Fig. 6 shows Fig. 1 removes the synoptic diagram of an embodiment of recognition template;
Fig. 7 shows the synoptic diagram be used to an embodiment of the method for going to identify the visual media data;
In whole instructions, similar label can be used for identifying similar parts.
Embodiment
Be readily appreciated that, here general description and in the accompanying drawings the parts of illustrated each embodiment can diversified different configuration be arranged and design.Therefore, be not to be intended to limit the scope of the present disclosure to the following detailed description as each embodiment of representing in the accompanying drawing, but only be used for each embodiment of explanation.Although the many aspects of embodiment present in the accompanying drawings, accompanying drawing is not necessarily drawn in proportion, unless specialize.
The present invention may be embodied in and not break away from other spirit or essential characteristic in other particular forms.It is illustrative that described embodiment will be understood that in all respects, and nonrestrictive.Therefore, scope of the present invention is indicated by appended claims, and can't help the indication of this detailed description.The implication and all changes within the scope that are in the equivalent of claim will be included within the scope of claim.
The feature of mentioning in the whole instructions, advantage or similar language do not mean that available all these feature and advantage of the invention process should or determine in single embodiment of the present invention.On the contrary, the language of mentioning feature and advantage is understood that to mean that being relevant to special characteristic, advantage or the characteristics that embodiment describes is included at least one embodiment of the present invention.Therefore, in the whole instructions to the discussion of characteristics and advantage and similar language throughout can but the identical embodiment of nonessential finger.
In addition, described feature of the present invention, advantage and characteristics can be combined by any way in one or more embodiments.Those skilled in the relevant art will understand, and according to the description here, the present invention can be implemented under the situation of the one or more special characteristics that do not have specific embodiment or advantage.In other cases, can identify in certain embodiments is not the accompanying drawing feature and advantage that appear among all embodiment of the present invention.
" embodiment ", " embodiment " or the similar language throughout of mentioning in the whole instructions means that the special characteristic, structure or the characteristics that are relevant to indicated embodiment description are included at least one embodiment of the present invention.Therefore, phrase " in one embodiment ", " in an embodiment " or the similar language throughout in the whole instructions can but nonessentially all refer to identical embodiment.
Although described a lot of embodiment here, at least some described embodiment have provided for the system and method that goes to identify from the secret or sensitive data of visual media data.More specifically, this system merges the image sequence from the visual media data acquisition, and automatically identifies the position that comprises text in the average image probably.The user can manually improve the selected part of the average image, to determine which partly comprises sensitivity or security information, and system generates template from the average image and selected part, this template can be applied to the image sequence in the described media data set, so that the security information in each image sequence of obfuscation.
Be used for determining going identification and so on some conventional methods to comprise at text data: (i) use edit tool hand labeled sensitizing range, and the zone in the solution documentation by Manual Logos so that deletion; (ii) go identification automatically, wherein use the text mining method to find the responsive fragment of text, such as title, date and address, no matter it is in structured data fields or in free text (freetext); And the (iii) fixedly known field in the structural data, for example, title, date and address can be removed by the code at this field structureization especially.Conventional method also manually marked region in case determine identification what.Be used for how going the conventional method identified to comprise removing the built-in code that is identified part or use method based on template.In the method based on template of routine, this template is to come manual creation by highlight the zone that (highlight) will write in the sample files of given classification.When the form (form) of limited quantity but the time spent, this method can be useful.
In image and video, sensitive information can embed in the various layouts, thereby creates a lot of Format Types.For DICOM image and video, the specific data of the patient who sees in image will change along with pattern (modality) type (echo or angiogram), manufacturer's (different manufacturers may show different information at its screen) and inspection itself (it further depends on corresponding disease and check result thereof) especially.The combination of these factors causes a large amount of commendation types that will see in the visual media data, become very loaded down with trivial details and the cost poor efficiency thereby manually learn template.For example, in typical echo record, nearly 50 different Format Types can occur, it is corresponding to about 146 detections that surpass of depending on disease by ultrasonic inspection person's record.Therefore, can at least semi-automatically generate system be used to the template of going to identify the visual media data can provide to various types of images and video fast and go identification efficiently.In certain embodiments, can be from the sample set study template of visual media data source.In addition, can realize correction or the modification of template by semi-automatic process described herein.In addition, the training stage forming template, template (or elementary version of template) can be applied to the candidate region in the image, so that the some or all of of identifying of picture/text content that define in the frame of recognition template may take place.
Fig. 1 shows the synoptic diagram that visual media removes an embodiment of recognition system 100.Shown visual media goes recognition system 100 to comprise the multiple assembly of following detailed description, and it can carry out function described herein and operation.In one embodiment, go at least some assemblies in the recognition system 100 in computer system, to realize.For example, go the function of one or more assemblies of recognition system 100 to realize by being stored in the computer memory device 100 and by the computer program instructions of carrying out such as the treatment facility 104 of CPU.Go recognition system 100 can comprise other assemblies, such as input-output apparatus 106, disk storage driver 108, image combiner 110, remove to identify engine 112 and template generator 114.Visual media goes some or all combinations of recognition system 100 can be stored in the single computing equipment, or is stored in the network of the computing equipment that comprises cordless communication network.Go recognition system 100 can comprise than more or less assembly described herein or subsystem.In certain embodiments, go recognition system 100 to can be used for implementing method as shown in Figure 7 described herein.
In one embodiment, image combiner 110 receives image sequence 116 from visual media data acquisition 118.In certain embodiments, described visual media data acquisition 118 can comprise image or the video that uses Medical Devices to catch, and comprises ultrasonoscopy, echocardiogram image, angiographic image or any other visual media.In other embodiments, described visual media data acquisition 118 can comprise image or the video that use is caught or generated for the equipment of other application.Each image sequence 116 can comprise a lot of images.For example, a video can comprise a lot of single image frames that per second is caught.
In certain embodiments, image combiner 110 is configured to receive image sequence 116 from a plurality of visual media data acquisitions 118.Each visual media data acquisition 118 can be corresponding to different Machine Types.Can a plurality of visual media data acquisitions be divided into groups according to predetermined classification, the categorizing system of described classification such as Machine Type, position, hospital, department or any other type, for described categorizing system, image sequence 116 in each visual media data acquisition 118 have with corresponding visual media data acquisition 118 in common some visual interest of other image sequences 116, such as layout, text similarity or other characteristics.
In one embodiment, image combiner 110 merges image in the image sequence 116 by come average pixel value 120 at each image of striding entire image sequence 116.Obtain the average image 122 like this, its pixel value 120 is that all images of striding in the image sequence 116 is average.In one embodiment, because the average image 122 comprises average pixel value 120, the average image 122 can only show all or the constant or fixing pixel of most of image of striding in the image sequence 116.The algorithm that is used for average pixel value 120 can be based on other characteristics of brightness, color value, saturation degree and/or pixel.In certain embodiments, the image of striding image sequence 116 pixel that do not satisfy preset frequency threshold value 124 is filtered out from the average image 122.This frequency threshold 124 can be any value of the text field in permission system 100 enough recognition image sequences 116.
In one embodiment, the pixel value 120 that produces in the average image 122 shows with the mean flow rate at the pixel value 120 of each location of pixels.Therefore, depend on the color mode of image, the pixel that does not have constant active value can be shown as black picture element in the average image 122.For example, comprise the black pixel value 120 of striding the constant pixel of all or most of image for not being from the average image 122 that is ultrasonic image sequence 116, have white or bright pixel value 120 and stride the pixel that all or most of image have a geostationary value.So any fixed text of image can be maintained in the average image 122, be constant because pixel value 120 is striden all images.
Removing to identify engine 112 is configured to define and is confirmed as relatively-stationary part in the average image 122.In one embodiment, be defined part 128 corresponding to the bright pixel value 120 in the average image 122.Remove to identify engine 112 and can keep to have a certain size or be positioned at and be defined zone 128 within other certain distances that are defined part 128, and discardable do not satisfy this requirement be defined zone 128.
In one embodiment, remove to identify engine 112 and be configured to define from the phase joined assemblies (component) of the average image 122 finding character 132 or the potential character in the average image 122, and produce character picture.Described phase joined assemblies can be comprising all pixels have the pixel region of bright pixel value 120.System 100 also can be configured to by using potential word 130 in optical character identification (OCR) the software identification the average image 122 and text to define word from equal image 122, to generate word image.Word 130 as described herein can comprise any combination of one or more characters 132.Based on word image and character picture, can keep and be defined part 128 so remove to identify engine 112, therein, being defined character 132 and overlapping from the word 132 that is defined of word image from character picture of certain predetermined percentage.Therefore, largely be defined the phase joined assemblies that is defined that word 130 or word components overlap and be maintained in the phrase image.
In certain embodiments, extract picture material and the content of text that is defined in the part.Then, identify text message and it is grouped in the semantic entity by the frame that defines of analyzing character and word.In other embodiments, can realize other operations and/or analyze and to extract and to identify content of text.
The visual media that Fig. 2 shows Fig. 1 removes the synoptic diagram of an embodiment of recognition system 100.The image sequence 116 that goes recognition system 100 to receive corresponding to visual media data acquisition 118.Go recognition system 100 to comprise image combiner 110, be used for and merge into the average image 122 from the image of image sequence 100.Image combiner 110 can use any method merging from the creation of image the average image of image sequence 116.Relative fixed or constant assembly in whole or some image in the average image displayable image sequence 116.
After the image in the image sequence 116 is merged into the average image 122, go to identify the part that engine 112 finds to comprise probably in the average image 122 text or word 130, and use the position of defining frame or other visual confining method retrtievals.Can import 200 based on the user improves and is defined part 128 to keep or to remove the frame that defines in the average image 122.
Fig. 3 illustrates for generating the process flow diagram of an embodiment that visual media goes the method for recognition template 126.Although method 300 is to describe in conjunction with the recognition system 100 of going of Fig. 1, the embodiment of method 300 can use the recognition system 100 of going of other types to implement.
In one embodiment, go recognition system 100 that image sequence 116 is merged into single the average image 122.For example, the average image 122 can anyly can be described this image sequence 116 in an image any way obtains, and this image has pixel value average on all images of this image sequence 116 120.Noise 304 also can be removed by this system 120 from the average image 122, more easily determine the part that will be defined of the average image 122 to allow recognition system 100.
The character picture that produces then and word image 312 can be used for finding to comprise in the average image 122 the most probable part of text.In one embodiment, go recognition system 100 to keep 314 parts that are defined that comprise one or more character frames of completely or partially overlapping with one or more word frames.For example, merge with the most of or portion of closing of a word frame or a plurality of character frames of overlapping can make the delimited area of recognition system 100 reservations corresponding to described character frame and word frame.The number percent of the coincidence between character frame and the word frame can allow draw characters image and word image 312 among both define frame 400 time some error, and still keep be defined the height possibility that the zone comprises text.In certain embodiments, go recognition system 100 to define phrase based on the distance between the delimited area.Can use by the delimited area that keeps the character/word frame generation that overlaps and create phrase image 316.This phrase image 316 can be used for generating the template 126 be used to removing to identify described image sequence 116 or other image sequences 116.
Fig. 4 shows the synoptic diagram of an embodiment of the average image 122 of Fig. 1.Although described in conjunction with the average image 122 of Fig. 4 here and gone recognition system 100, gone recognition system 100 to use in conjunction with any the average image 122.
In one embodiment, remove the phase joined assemblies in the recognition system 100 discovery the average images 122, and define frame 400 around the drafting of phase joined assemblies.The phase joined assemblies can be that the text 132 that links to each other in one group, image artifacts (artifact) or other have a plurality of zones of enlivening pixel region.Character defines algorithm can define frame 400 around drawing greater than pre-sizing or the assembly that links to each other that satisfies certain other threshold values, and this can help to reduce the error that detects the character 132 in the average image 122.Do not satisfy assembly or the pixel of threshold value in the average image 122 and can stay in the average image 122, maybe can from image, remove to remove noise or unnecessary assembly.Can be by indicating clear visual separation between the phase joined assemblies in the average image 122 defining leaving space between the frame 400 around the frame 400 that defines that the assembly that links to each other is drawn.
In one embodiment, during phase joined assemblies in defining the average image 122, the user can manually improve the position of defining frame 400 or the quantity in the character picture 308.The user can remove In the view of the user obviously be non-character element define frame 400.Perhaps, the user can be around the phase joined assemblies hand drawn frame that goes recognition system 100 to miss.
Fig. 5 shows the synoptic diagram of an embodiment of the average image 122 of Fig. 1.Although described in conjunction with the average image 122 of Fig. 5 here and gone recognition system 100, gone recognition system 100 to use in conjunction with any the average image 122.
In one embodiment, go recognition system 100 to find word 130 in the average images 122, and draw around word 130 or the zone that comprises text probably and to define frame 400.It is text filed to go recognition system 100 can use the OCR engine to find.The OCR engine can determine that each zone in the average image 122 has the value of the confidence of text.If the value of the confidence satisfies certain threshold value, this zone is considered to the candidate, draw for this zone and define frame 400, and produce respectively define the word image 312 that frame 400 is configured for the average image 122.
In one embodiment, during text filed in defining the average image 122, the user can manually improve the position of defining frame 400 or the quantity in the word image 312 that produces.The user can remove In the view of the user obviously be non-word or text component define frame 400.Perhaps, the user can be the regional hand drawn frame that recognition system 100 is missed.
Fig. 6 shows the synoptic diagram that visual media removes an embodiment of recognition template 126.Although described in conjunction with the average image 122 of Fig. 6 here and gone recognition system 100, gone recognition system 100 to use in conjunction with any the average image 122.
In certain embodiments, frame 400 is defined in the assembly that is not defined in character picture 308 drafting of going recognition system 100 can center in the word image 312.Otherwise frame 400 is defined in the assembly that does not define in character picture 312 drafting of going recognition system 100 can center in the character picture 308.By compare string image 308 and word image 312, go recognition system 100 can determine which frame most possibly comprises text.
Can be used for generation then by the delimited area of going recognition system 100 to keep and remove recognition template 126.The content that is defined the position 134 definable templates of zone in phrase image 316 will be by the part of obfuscation, and described phrase image 316 can be the average image 122, has wherein drawn in some zone of the average image 122 and has defined frame 400.When template 126 is applied to image in the image sequence 116, go recognition system 100 can seek in the image zone corresponding to the position that is defined zone 128 of template 126, and these zones in the obfuscation image automatically.In certain embodiments, template 126 can be applied to Streaming Media, thereby can use template 126 to handle the live video record, to remove sensitive data from live video.
Fig. 7 shows the process flow diagram be used to an embodiment of the method 700 of going to identify the visual media data.Although the recognition system 100 of going in conjunction with Fig. 1 has been described method 700, the embodiment of method 700 can use the recognition system 100 of going of other types to realize.
In one embodiment, go recognition system 100 to merge 710 from the image sequence 116 of visual media data acquisition 118 and be the average image 122.The average image 122 can be created by striding average 705 pixel values 120 of all images in the image sequence 116 and filter out 715 pixel values 120 that do not satisfy preset frequency threshold value 124 from the average image 122 by system 100.
In one embodiment, system 100 defines the relatively-stationary part that is confirmed as of 720 the average images 122.Each is defined part 128 and can be identified by the relevant position 134 in the average image 122.In one embodiment, the part that defines the average image 122 comprises defining from the phase joined assemblies of the average image 122 and finds that character 132 is to produce character picture 308.The part that defines the average image 122 comprises and defining from the word of the average image 122 to produce word image 312.This can comprise that the part of analyzing the average image 122 comprises the value of the confidence of text with the part that obtains to analyze.The part of analyzing can be analyzed by the OCR engine.In one embodiment, this OCR engine is adapted to especially with going recognition system 100 to use.System 100 can satisfy the word threshold value in response to definite the value of the confidence, determines that the part of analyzing is word candidates.System 100 can keep such part 128 that is defined then, wherein being defined character and overlapping from the word 130 that is defined of word image 312 from character picture 308 of predetermined percentage.In one embodiment, system 100 merges to be positioned at mutually and is defined part 128 to form phrase within the predetermined mean distance.A plurality of phrases can be defined at together, so that it is more efficient to be defined the obfuscation in zone.
Then, can use 735 templates 126 going to identify for the image sequence 116 that generates template 126, this is to realize by the content corresponding to the position that is defined part 128 of template 126 in the obfuscation image.This all images that can allow to stride in the image sequence 116 carries out consistent obfuscation.In one embodiment, template 116 is used to identify other image sequences 116 in the visual media data acquisition 118 then.Described other image sequences 116 can be shared adjacent characteristics with the image sequence 116 that is used for generation template 126, such as the placement position of the text in each image, object or other assemblies.In certain embodiments, system 100 can be each visual media data acquisition 118 and generates a template 126, thereby system 100 can go recognition image sequence 116 for each different viewdatas set 118 automatically.
Although system 100 given here and method are relevant to and identify the visual media data and describe, but system's 100 methods also can be used for identifying the data of text data or other types.
The embodiment of phrase match system 100 comprises at least one processor that directly or indirectly is coupled by system bus and memory component such as data, address and/or control bus.The local storage, mass storage and the cache memory that use the term of execution that memory component can being included in program code actual, cache memory provides the interim storage of some program code at least, so as to reduce in the process of implementation must be from mass storage the number of times of retrieval coding
Should also be noted that to use for some operation at least of described method and be stored in computer-usable storage medium so that the software instruction of being carried out by computing machine is realized.As example, the embodiment of computer program comprises the computer-usable storage medium of storing computer-readable program, when this computer-readable program is carried out on computers, make to comprise this computing machine executable operations from the operation of electronic document information extraction.
Although the operation of the method here is shown and described with particular order, the order of the operation of each method can be changed, thus some operation can reverse order carry out, thereby perhaps some operation can be carried out with other operations at least in part concomitantly.In another embodiment, the instruction of different operating or child-operation can be intermittently and/or the mode that replaces implement.
The present invention can take complete hardware embodiment, complete software embodiment or comprise the form of the embodiment of software and hardware both elements.In one embodiment, the present invention realizes that in software described software includes but not limited to firmware, resident software, microcode etc.
In addition, embodiments of the invention can take can by computing machine can with or the form of the computer program of computer-readable medium visit, it provides by computing machine or any instruction execution system and has carried out or the program code of carrying out associated with it.For this illustrative purposes, computing machine can with or computer-readable medium can be can comprise, storage, transmission, propagation or transmission procedure be in order to carried out or any device of carrying out associated with it by instruction execution system, device or equipment.
Computing machine can use or computer-readable medium can be electronics, magnetic, light, electromagnetism, infrared or semiconductor system (or device or equipment), or is propagation medium.The example of computer-readable medium comprises semiconductor or solid-state memory, tape, dismantled and assembled computer format floppy, random-access memory (RAM), ROM (read-only memory) (ROM), rigid magnetic disks and CD.The current example of CD comprises compact disc read-only memory (CD-ROM), read-write CD (CD-RW) and digital video disc (DVD).
I/O or I/O equipment (including but not limited to keyboard, display, pointing apparatus) can be directly or the I/O controller by the centre be connected to system.In addition, network adapter also can be connected to system, so that data handling system can be connected to other data handling systems or remote printer or memory device etc. by special use or the public network of centre.Modulator-demodular unit, cable modem and Ethernet card only are several examples of current available network type of adapter.
The specific detail of a plurality of embodiment is provided in the above description.Yet some embodiment can be with being implemented, and be less than all these details.In other cases, for concise and to the point and clear for the purpose of, description the level of detail of some method, process, assembly, structure and/or function is limited to realize a plurality of embodiment of the present invention.
Although described and shown specific embodiment of the present invention, the present invention is not limited to particular form or the arrangement of the parts so describing and illustrate.Scope of the present invention will be limited by claims and equivalent thereof.
Claims (13)
1. method that is used for going to identify the visual media data comprises:
To merge into the average image from the image sequence of visual media data acquisition;
Define the relatively-stationary part that is confirmed as of the average image, wherein each is defined partly and is identified by the relevant position in the average image;
Generation comprises that each that be defined part and be used for the average image is defined the template of the relevant position of part; And
By obfuscation be defined the part content make a return journey identification described image sequence.
2. the process of claim 1 wherein, merge image sequence and also comprise:
The all images of striding in the image sequence comes average pixel value, to obtain the average image; And
From the average image, filter out the pixel value that does not satisfy the preset frequency threshold value.
3. the method for claim 1 also comprises:
Applying template is to remove to identify other image sequences in the described visual media data acquisition.
4. the process of claim 1 wherein that the relatively-stationary part that is confirmed as that defines the average image also comprises:
Define from the phase joined assemblies of the average image to find character and to produce character picture;
Define from the word of the average image to produce word image; And
Keep to be defined part, described being defined character and overlapping from the word that is defined of word image from character picture that is defined predetermined percentage in the part.
5. the method for claim 4, wherein, the word that defines from the average image also comprises:
Analyze the part of the average image to obtain the degree of confidence that analyzed part comprises text; And
Satisfy the word threshold value in response to definite degree of confidence, determine that the part of analyzing is word candidates.
6. the method for claim 1 also comprises:
Merge and be positioned at being defined partly to form phrase within the predetermined mean distance mutually.
7. the method for claim 1 also comprises:
Import the part that is defined of improving the average image based on the user; And
Import the authority of setting up for being defined part of checking from the user, wherein, the described authority of checking determines when that being defined content partly can check for given user.
8. a visual media goes recognition system, comprising:
The image combiner, it is configured to and will merges into the average image from the image sequence of visual media data acquisition; And
Remove to identify engine, it is configured to:
Define the relatively-stationary part that is confirmed as of the average image, wherein each is defined partly and is identified by the relevant position in the average image;
Generation comprises that each that be defined part and be used for the average image is defined the template of the relevant position of part; And
By obfuscation be defined the part content make a return journey identification described image sequence.
9. the system of claim 8, wherein, described image combiner also is configured to:
The all images of striding in the image sequence comes average pixel value, to obtain the average image; And
From the average image, filter out the pixel value that does not satisfy the preset frequency threshold value.
10. the system of claim 8 also comprises template generator, and it is configured to:
Applying template is to remove to identify other image sequences in the described visual media data acquisition.
11. the system of claim 8, wherein, the relatively-stationary part that is confirmed as that defines the average image also comprises:
Define from the phase joined assemblies of the average image to find character and to produce character picture;
Define from the word of the average image to produce word image; And
Keep to be defined part, described being defined character and overlapping from the word that is defined of word image from character picture that is defined predetermined percentage in the part.
12. the system of claim 11, wherein, the word that defines from the average image also comprises:
Analyze the part of the average image to obtain the degree of confidence that analyzed part comprises text; And
Satisfy the word threshold value in response to definite degree of confidence, determine that the part of analyzing is word candidates.
13. the system of claim 8 wherein, describedly removes to identify engine and also is configured to:
Import the part that is defined of improving the average image based on the user.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US13/351,141 US9147178B2 (en) | 2012-01-16 | 2012-01-16 | De-identification in visual media data |
| US13/351,141 | 2012-01-16 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103294746A true CN103294746A (en) | 2013-09-11 |
| CN103294746B CN103294746B (en) | 2017-07-18 |
Family
ID=48779652
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310016358.9A Active CN103294746B (en) | 2012-01-16 | 2013-01-16 | For the method and system for going identification in visual media data |
Country Status (2)
| Country | Link |
|---|---|
| US (2) | US9147178B2 (en) |
| CN (1) | CN103294746B (en) |
Families Citing this family (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101508822B1 (en) | 2008-06-02 | 2015-04-07 | 프리시젼 바이오메트릭스, 인크. | Systems and methods for performing surface electromyography and range-of-motion tests |
| US11557073B2 (en) * | 2008-06-02 | 2023-01-17 | Precision Biometrics, Inc. | System for generating medical diagnostic images |
| US8793199B2 (en) * | 2012-02-29 | 2014-07-29 | International Business Machines Corporation | Extraction of information from clinical reports |
| US8923647B2 (en) * | 2012-09-25 | 2014-12-30 | Google, Inc. | Providing privacy in a social network system |
| CA2852253A1 (en) * | 2014-05-23 | 2015-11-23 | University Of Ottawa | System and method for shifting dates in the de-identification of datesets |
| US20160203264A1 (en) * | 2015-01-09 | 2016-07-14 | Intelemage, Llc | Systems, methods, and computer program products for processing medical images to address privacy concerns |
| US20160307063A1 (en) * | 2015-04-16 | 2016-10-20 | Synaptive Medical (Barbados) Inc. | Dicom de-identification system and method |
| US10169548B2 (en) | 2016-08-24 | 2019-01-01 | International Business Machines Corporation | Image obfuscation |
| US10860985B2 (en) | 2016-10-11 | 2020-12-08 | Ricoh Company, Ltd. | Post-meeting processing using artificial intelligence |
| US11307735B2 (en) | 2016-10-11 | 2022-04-19 | Ricoh Company, Ltd. | Creating agendas for electronic meetings using artificial intelligence |
| US10572858B2 (en) | 2016-10-11 | 2020-02-25 | Ricoh Company, Ltd. | Managing electronic meetings using artificial intelligence and meeting rules templates |
| US10510051B2 (en) | 2016-10-11 | 2019-12-17 | Ricoh Company, Ltd. | Real-time (intra-meeting) processing using artificial intelligence |
| US10250592B2 (en) | 2016-12-19 | 2019-04-02 | Ricoh Company, Ltd. | Approach for accessing third-party content collaboration services on interactive whiteboard appliances using cross-license authentication |
| US10375130B2 (en) | 2016-12-19 | 2019-08-06 | Ricoh Company, Ltd. | Approach for accessing third-party content collaboration services on interactive whiteboard appliances by an application using a wrapper application program interface |
| US10298635B2 (en) | 2016-12-19 | 2019-05-21 | Ricoh Company, Ltd. | Approach for accessing third-party content collaboration services on interactive whiteboard appliances using a wrapper application program interface |
| US10395405B2 (en) | 2017-02-28 | 2019-08-27 | Ricoh Company, Ltd. | Removing identifying information from image data on computing devices using markers |
| US11138778B2 (en) | 2017-06-29 | 2021-10-05 | Koninklijke Philips N.V. | Obscuring facial features of a subject in an image |
| US10552546B2 (en) | 2017-10-09 | 2020-02-04 | Ricoh Company, Ltd. | Speech-to-text conversion for interactive whiteboard appliances in multi-language electronic meetings |
| US11030585B2 (en) | 2017-10-09 | 2021-06-08 | Ricoh Company, Ltd. | Person detection, person identification and meeting start for interactive whiteboard appliances |
| US10956875B2 (en) | 2017-10-09 | 2021-03-23 | Ricoh Company, Ltd. | Attendance tracking, presentation files, meeting services and agenda extraction for interactive whiteboard appliances |
| US10553208B2 (en) | 2017-10-09 | 2020-02-04 | Ricoh Company, Ltd. | Speech-to-text conversion for interactive whiteboard appliances using multiple services |
| US11062271B2 (en) | 2017-10-09 | 2021-07-13 | Ricoh Company, Ltd. | Interactive whiteboard appliances with learning capabilities |
| US10757148B2 (en) | 2018-03-02 | 2020-08-25 | Ricoh Company, Ltd. | Conducting electronic meetings over computer networks using interactive whiteboard appliances and mobile devices |
| EP3847657A1 (en) | 2018-09-05 | 2021-07-14 | Translational Imaging Innovations LLC | Methods, systems and computer program products for retrospective data mining |
| US11080424B2 (en) * | 2019-05-21 | 2021-08-03 | Verb Surgical Inc. | Method and system for anonymizing raw surgical procedure videos |
| US11606336B2 (en) * | 2020-05-18 | 2023-03-14 | Lynx Md Ltd | Determining permissions in privacy firewalls |
| US11652721B2 (en) * | 2021-06-30 | 2023-05-16 | Capital One Services, Llc | Secure and privacy aware monitoring with dynamic resiliency for distributed systems |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040193901A1 (en) * | 2003-03-27 | 2004-09-30 | Ge Medical Systems Global Company, Llc | Dynamic configuration of patient tags and masking types while de-identifying patient data during image export from PACS diagnostic workstation |
| US20080046757A1 (en) * | 2006-07-12 | 2008-02-21 | Palo Alto Research Center Incorporated | Method, Apparatus, and Program Product for Flexible Redaction of Content |
| CN101678404A (en) * | 2007-05-30 | 2010-03-24 | 索利斯蒂克有限公司 | Method of handling transmittals including a graphic classification of the signatures associated with the transmittals |
| CN101888469A (en) * | 2009-05-13 | 2010-11-17 | 富士通株式会社 | Image processing method and image processing device |
| US7970213B1 (en) * | 2007-05-21 | 2011-06-28 | A9.Com, Inc. | Method and system for improving the recognition of text in an image |
| CN102301376A (en) * | 2008-12-23 | 2011-12-28 | 克洛西克斯解决方案公司 | Double blinded privacy-safe distributed data mining protocol |
Family Cites Families (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5442715A (en) * | 1992-04-06 | 1995-08-15 | Eastman Kodak Company | Method and apparatus for cursive script recognition |
| US6021220A (en) * | 1997-02-11 | 2000-02-01 | Silicon Biology, Inc. | System and method for pattern recognition |
| WO2002006948A1 (en) | 2000-07-13 | 2002-01-24 | Digineer, Inc. | Method for protecting the privacy, security, and integrity of sensitive data |
| US7519591B2 (en) | 2003-03-12 | 2009-04-14 | Siemens Medical Solutions Usa, Inc. | Systems and methods for encryption-based de-identification of protected health information |
| EP1555804A3 (en) * | 2004-01-19 | 2006-08-16 | Ricoh Company, Ltd. | Image processing apparatus, image processing program and storage medium |
| US20060081714A1 (en) * | 2004-08-23 | 2006-04-20 | King Martin T | Portable scanning device |
| BRPI0520469A2 (en) | 2005-07-29 | 2009-06-13 | Telecom Italia Spa | automatic biometric identification method and system, and, computer program product |
| CN101464951B (en) | 2007-12-21 | 2012-05-30 | 北大方正集团有限公司 | Image recognition method and system |
| US8270718B2 (en) * | 2008-09-23 | 2012-09-18 | International Business Machines Corporation | Manipulating an image by applying a de-identification process |
| US8515185B2 (en) * | 2009-11-25 | 2013-08-20 | Google Inc. | On-screen guideline-based selective text recognition |
| US20120033892A1 (en) * | 2010-08-04 | 2012-02-09 | Coreguard | Systems and Methods for Identifying Matching Images of Digital Documents |
| US8285074B2 (en) * | 2010-09-01 | 2012-10-09 | Palo Alto Research Center Incorporated | Finding low variance regions in document images for generating image anchor templates for content anchoring, data extraction, and document classification |
| JP5144736B2 (en) * | 2010-11-10 | 2013-02-13 | シャープ株式会社 | Document generation apparatus, document generation method, computer program, and recording medium |
-
2012
- 2012-01-16 US US13/351,141 patent/US9147178B2/en active Active
- 2012-08-28 US US13/597,085 patent/US9147179B2/en not_active Expired - Fee Related
-
2013
- 2013-01-16 CN CN201310016358.9A patent/CN103294746B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040193901A1 (en) * | 2003-03-27 | 2004-09-30 | Ge Medical Systems Global Company, Llc | Dynamic configuration of patient tags and masking types while de-identifying patient data during image export from PACS diagnostic workstation |
| US20080046757A1 (en) * | 2006-07-12 | 2008-02-21 | Palo Alto Research Center Incorporated | Method, Apparatus, and Program Product for Flexible Redaction of Content |
| US7970213B1 (en) * | 2007-05-21 | 2011-06-28 | A9.Com, Inc. | Method and system for improving the recognition of text in an image |
| CN101678404A (en) * | 2007-05-30 | 2010-03-24 | 索利斯蒂克有限公司 | Method of handling transmittals including a graphic classification of the signatures associated with the transmittals |
| CN102301376A (en) * | 2008-12-23 | 2011-12-28 | 克洛西克斯解决方案公司 | Double blinded privacy-safe distributed data mining protocol |
| CN101888469A (en) * | 2009-05-13 | 2010-11-17 | 富士通株式会社 | Image processing method and image processing device |
Also Published As
| Publication number | Publication date |
|---|---|
| US20130182007A1 (en) | 2013-07-18 |
| CN103294746B (en) | 2017-07-18 |
| US9147179B2 (en) | 2015-09-29 |
| US20130182006A1 (en) | 2013-07-18 |
| US9147178B2 (en) | 2015-09-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103294746A (en) | Method and device of de-identification in visual media data | |
| CN107239666B (en) | Method and system for desensitizing medical image data | |
| US10216958B2 (en) | Minimizing sensitive data exposure during preparation of redacted documents | |
| JP4676225B2 (en) | Method and apparatus for capturing electronic forms from scanned documents | |
| JP6832867B2 (en) | Methods and devices for verifying images based on image verification codes | |
| US10068380B2 (en) | Methods and systems for generating virtual reality environments from electronic documents | |
| US9754120B2 (en) | Document redaction with data retention | |
| US9870484B2 (en) | Document redaction | |
| CN110956026B (en) | Legal document generation method and device and electronic equipment | |
| JP2006260318A (en) | Diagnostic reading report input support method and system | |
| US10290365B2 (en) | Image processing apparatus, image processing method, and non-transitory computer readable medium | |
| US20160140145A1 (en) | Extracting information from PDF Documents using Black-Box Image Processing | |
| US20140297269A1 (en) | Associating parts of a document based on semantic similarity | |
| CN113853658A (en) | Method and system for anonymizing original surgical procedure video | |
| Chiatti et al. | Text extraction and retrieval from smartphone screenshots: Building a repository for life in media | |
| US20160085731A1 (en) | Reordering Text from Unstructured Sources to Intended Reading Flow | |
| US20160275245A1 (en) | Iterative construction of clinical history sections | |
| JP6529254B2 (en) | INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, PROGRAM, AND STORAGE MEDIUM | |
| Chu et al. | Multimodal retrieval through relations between subjects and objects in lifelog images | |
| CN109299214B (en) | Text information extraction method, text information extraction device, text information extraction medium and electronic equipment | |
| JP2009217428A (en) | Translation display, translation display method, and translation display program | |
| Tsui et al. | Automatic selective removal of embedded patient information from image content of DICOM files | |
| CN116758565B (en) | OCR text restoration method, equipment and storage medium based on decision tree | |
| RU2413985C2 (en) | Method of converting weakly-formalised documents in order to minimise volume thereof during storage | |
| JP6132422B2 (en) | Old electronic medical record data activation system for computer with new electronic medical record system program |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |